Among the more challenging and complex public health problems today, college alcohol use correlates with problems such as academic underperformance, dropping out, injury, death, and high-risk sexual behavior. Despite serious effort and expense, the factors driving alcohol usage or effective means of intervention remain uncertain. Out-of-the-box thinking is required for the design of effective intervention approaches. Theoretical and computational models, properly constructed, can provide important insights. Agent-based models can illuminate the importance of social networks and peer groups, significant differences among individuals, and detailed campus environmental variables, such as parties, athletic events, that are difficult to examine with deterministic compartmental models. Individual or agent-based models offer attractive supplemental tools that can be used to make predictions as well as to inform the validation and model enhancement efforts for compartmental models.